I can't imagine myself using reflection much, but maybe it will eliminate a lot of feature proposals bogging down the committee and they can focus on harder problems.
It would be cool if the stated goal of C++29 was compile times.
I'd argue reflection is very much a feature for libraries. You wouldn't use it directly, but your JSON / YAML serialize is then built on top of it. So are your bindings for scripting engines like Lua.
You can already automatically serialize/deserialize arbitrarily nested structs since C++17 (using Boost.PFR). Since C++20, you can also serialize/deserialize the struct data member names automatically.
There are a lot of things that are very very important for a tiny niche. In any non-trivial project you will end up with a lot of custom libraries and some of them really benefit from some obscure feature that no place else in your project would want.
Also nice for UI tooling; game tools, debuggers, etc. Pull apart a struct and display it on screen and not have to patch the UI tool every time you change the struct is pretty nice.
It would be cool if the stated goal of C++29 was compile times.